Babel Street Text Analytics is a powerful software solution designed to provide multilingual capabilities for government and commercial organizations. It leverages advanced natural language processing (NLP) and artificial intelligence (AI) techniques to analyze unstructured text data in over 40 languages. The software addresses the challenges of working with multilingual data by offering accurate and comprehensive language analysis, enabling more informed decision-making on a global scale. Key functionalities include precise language identification and text encoding, enabling language-specific analysis. The software performs tasks such as morphological analysis, character normalization, tokenization, part-of-speech tagging, lemmatization, noun decompounding, and pronunciation analysis for Japanese and Chinese. It excels at entity extraction, identifying nearly 20 entity types (including people, organizations, and locations) from unstructured text, distinguishing between similar entities and linking them to knowledge base entries. This allows users to understand the "who, what, when, and where" of events and relationships, significantly reducing analyst workload. Furthermore, Babel Street Text Analytics offers cross-lingual search capabilities powered by semantic similarity. This allows users to search across multiple languages using key phrases, retrieving highly relevant results. Search results can be filtered by various criteria, such as people, organizations, locations, and sentiment. The software also provides tools to understand concepts and key phrases within the text, facilitating efficient information retrieval. An Adaptation Studio allows users to train and fine-tune models for domain-specific entities and events, using active and few-shot learning techniques to reduce model training time without requiring coding.

Features

Unmatched Linguistic Analysis

• Provides accurate and comprehensive language analysis across 40+ languages.
  • Morphological analysis
  • Character normalization
  • Tokenization
  • Part-of-speech tagging
  • Lemmatization
  • Noun decompounding
  • Pronunciation analysis (Japanese and Chinese)

Entity Extraction and AI Text Analysis

• Extracts nearly 20 entity types, distinguishes between similar entities, and links them to knowledge base entries.

Cross-lingual Search

• Enables semantic similarity-powered search across multiple languages using key phrases.

NLP-enabled Search Enhancements

• Allows filtering search results by people, organizations, location, and sentiment; provides concept, key phrase, and translated name views.

Train, Adapt Models (Adaptation Studio)

• Improves accuracy by training and fine-tuning models for domain-specific entities and events using active and few-shot learning (no coding required).